Water Yam (Dioscorea alata L.) Growth and Tuber Yield as Affected by Rotation and Fertilization Regimes across an Environmental Gradient in West Africa

Round 1

Reviewer 1 Report

Dear authors, I have read with interest your manuscript "Water yam (Dioscorea alata L.) growth and tuber yield as affected by rotation and fertilization regimes across an environmental gradient in West Africa". The topic is important for food security in West Africa especially as publications on the subject are rare. The main strength of the article is the longitudinal gradient covered by the study that is used to explain yam yield main driver. The entire article is well written and all the ideas are explained and connected with international literature.

 

General remarks:

With such a complex experimental design and so many explanatory variables, statistics could have been oriented towards methods dedicated to multivariate analysis such as path analysis or Bayesian networks. Using mixed model, it is not clear to me which equation was used to test which hypothesis. At least authors should detail the equations, and summarize mixed model results (e.g. fixed effect estimates with confidence intervals and random effect with standard deviation and p-value).

The choice of site variables in the different analyses is not clear to me. Why is pH and SOC included in the PCA while not P or N for instance? How was SOC identified as a determinant among others (e.g. N content, pH...)? Why rainfall variables was not included into PCA?

 

Specifics remarks:

Positioning developmental phases into Figs 1-3 could help the reader to interpret graphs.

A map illustrating experimental sites in supplementary material could help to remember sites position and characteristics.

L184-201: Understanding how much nutrient was applied is not straightforward if we do not know yam exportation (e.g. kg N per t of fresh yield). Moreover, Table 1 increases confusion. If ORG and MIN are calculated on the basis of the same nitrogen exports, how can ORG provide almost 3 times more nitrogen than MIN (even if only 60% is estimated to be usable)?

L271-273: Authors should precise that despite good correlation, wooden frame method overestimate SSC by about 30% compared to canopeo (Fig S2). So merging both 2016 and 2018 SSC values within the same analysis (Fig S5) should be taken with cautions. Maybe the calculation of a relative SSC for each year (SSC/yearly max(SSC)) could allow to improve analysis results interpretation.

L284-285: How was DTY calculated (oven dried samples or theoretically derived estimation)?

L344-345: Not clear which type of analysis lead to this statement (mixed model, PCA, correlation?).

L530: SSC_70 and not SC_70 (idem for 98, 126 and 154)

L539: Both year PCA exhibit a negative relationship between soil pH and final emergence. Could this be linked with soil pathogens populations (no pest treatment in 2018)?

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Thank you for this interesting manuscript. I have some minor questions and recommendations as follows:

General questions

• Which long-term average of precipitaion characterizes each respective site?

• Which precrops were planted before the beginning of the experiment (2015)?

• Were the replicates randomised? Please add this information to the materials and methods section.

• Table 1: Why was so much more N applied in Liliyo and Tiéningboué for ORG compared to MIN?

• Table S1: Why was so much N applied to maize in ORG in Liliyo (more than 700 kg)? This value is not realistic. Please check, if there is any mistake in the calculation of the amount of N applied with poultry manure in Liliyo.

 

Inconsitencies

• Figure 1: The description of figure 1says that this figure shows precipiation from 14 days before planting until senescen. However, the x-axis starts with 0 days after planting. It should start with -14 (14 days before planting).

• Line 408-411: “In 2016, the number of tubers of 16,510 and 18,980 tubers ha⁻¹ harvested in Liliyo and Léo respectively, were statistically similar but significantly lower compared to that of 24,470 tubers ha⁻¹ harvested in Léo and statistically higher in comparison to that of 12,830 tubers ha⁻¹ harvested in Tiéningboué (Figure S3).” Looking at the boxplots (Figure S3) it seems that the number of tubers was highest in Midebdo, not in Léo.

• Figure S4: “The regression equations, the coefficients of determination and the p values are shown on the graphs.” Figure S4 only consists of one graph (not several graphs) and the mentioned parameters are not shown.

• Figure S5: “The linear regression, the coefficient of determination (r2) and the p-value are shown in the graph.” None of these parameters are shown.

• Table S6: “The values in brackets represent the coefficients of variation. The values in parenthesis represent the coefficients of variation expressed in % of the mean yield.” These two sentences have the same meaning. Please delete one of them.

   

 

Spelling mistakes

• Line 420/421: Please replace “variables” by “variable”.

• Line 513: Please replace “cluster 2 bring” by “cluster 2 brings”.

• Line 526: Please replace “YamLer “by “YamLeg”.

• Line 613: Please repalce “fertilized” by “fertilizer”.

• Supplementary material, line 118/119: “Each point of represent a mean os 192 observation[…]”. Please replace this sentence by “Each point represents a mean of 192 observations […]”.

 

Author Response

Please see the attachment.

Author Response File: Author Response.pdf